News for 2008

Thanks to our MIT contacts (and to satisfy popular demand) we will now have a TWO DAY conference, on March 27 and 28, 2008 , (Thursday and Friday) on campus at MIT in Cambridge, Massachusetts, USA.

On Thursday, we will be in the Stata Center (Building 32), Room 141(map), and on Friday, we will be in the next building down, Building 34, room 101 (the large hall) (map) .

Directions to the Site

The easy algorithm for getting there is to get on the T (the Boston Subway), get out at the Red Line Kendall Square station and stand in the middle of the street. Look around; you will see the Kendall Mariott hotel (big tall hotel). Walk past the hotel with the hotel on your right side (this is west). Walk two blocks; you will be at a funny intersection with the road to the left angled at about 45 degrees. Take the 45-degree road. The first building on your left is building 32- the Stata center; it looks like something only Gehry could design. Building 34 is another 100 meters further along, also on the left.

This two-day format will allow us time to add workshops and tutorial sessions, and give us additional flexibility in accepting papers. Our scientific standards won't change but we will be able to accomodate more information and different styles of presentation than the prior 20-minutes-of-Powerpoint system.

We will also have an evening "beer and munchies" event hosted by the ISP division of Comcast, on Thursday evening at a local, (and very high quality) "establishment". Details will be distributed at the afternoon session on Thursday.

Registering for the 2008 MIT Spam Conference

To register, just send an email to . Note that this mail will be read by a real human, so free-form formatting is fine and we'll get a response back to you. Note - this *will* put you on the human-administered conference mailing list. We appreciate preregistering; that way, we know how much food and drink we should have on hand.

2008 Papers now available for download!!!

The gzipped tar archive of the MIT Spam Conference 2008 is now available for download:

Individual papers & slide presentations are linked from the "Accepted Papers" list below.

Accepted Papers

We now have an accepted papers list. Here's the authors, affiliations, and papers:
  • A. O'Donnell - Cloudmark - Game Theory applied to Zombie Bots (keynote!)
  • J. Aycock - U. Calgary - Black Market Botnets
  • B. Bruen - KnujOn - KnujOn
  • A. Cosoi - BitDefender - False-positive-safe Neural Networks
  • S. Kubisch - Rostock U. - IPclip
  • M. O'Reirdan - ComCast - ISP Botfighting
  • K. Liszka - U. Akron - Neural Networks for Image Spam
  • A. McElligott - Geobytes - Casekeys
  • C. Mills - Microsoft - Metafeature training
  • J. Nagle - SiteTruth - Blacklisting Collateral Damage
  • P. Tom - - Botnet Clusters
  • R. Rajabiun - ComdomSoft - Multilayer Filtering (PDF of paper)
  • S. Sarafijanovic - EPFL - Open Digests
  • N. Yamai - Okayama U. - SMTP Session Abort
  • J. Zdziarski - Secure Computing Corp. - Adaptive Parsing
  • Additionally, we will have workshops on the following topics - these are open-ended "guided discussions", typically for one hour:

  • A. Cosoi - Social Networks and their implications
  • J. Fitzpatrick - the Malware Ecosystem
  • T. Eggendorfer - Introduction to Advanced Filtering
  • W. Yerazunis - Best Practices: What's in YOUR Spam Firewall?
  • The Chair wishes to thank the referees (who remain nameless), who by their care and consideration helped produce these high quality papers.

    Tentative Agenda

    Thursday Day
    Time Item Who Topic
    09:20AM Breakfast EveryoneCoffee, Bagels, Juice, Cake
    09:50AM Chair Opening Yerazunis
    10:05AM Keynote O'DonnellGame Theory of Botnets
    10:35AM Botnet Track AycockBlack Market Botnets
    10:55AM TomBotnet Clusters
    11:15AM O'Rierdan Major ISP Botnetfighting
    11:35AM Lunch
    01:05PM Workshop 1Fitzpatrick The Malware Ecosystem
    02:05PM **Break**
    02:15PM SMTP Track BruenKnujOn
    02:35PM YamaiSMTP Session Abort
    02:55PM KubischIPClip
    03:15PM **Break**
    03:15PM RajabiunMultilayer Filtering
    03:35PM McElligottCaseKeys
    03:55PM Workshop 2EggendorferAdvanced Filtering
    04:55PM Adjourn Reception
    Thursday Night
    06:00 PMReception Courtesy COMCAST
    09:00 AMBreakfast EveryoneCoffee, Bagels, Juice, Cake
    09:30 AMChair OpeningYerazunis
    09:40 AM Neural Networks Cosoi False-positive-safe Neural Networks
    10:00 AM Liszka Neural Networks for Image Spam
    10:20 AM Mills Metafeature Training
    10:40 AM **Break**
    10:50 AM Workshop 3 Cosoi Social Networks
    11:50 AM Lunch
    01:20 PM Advanced Filtering Sarafijanovic Open Digests
    01:40 PM Nagle Blacklisting Collateral Damage
    02:00 PM Zdziarski Adaptive Parsing
    02:20 PM **Break**
    02:30 PM Workshop 4 Yerazunis What's in YOUR Spam Filter?
    03:30 PM Best Papers Vote Everybody
    03:40 PM Closing Remarks All authors/leads
    03:50 PM Adjourn Yerazunis

    New flexibility in paper submissions

    The new longer conference, plus experience in last year's issues with the U.S. State Department delays in issuing entry visas, leads us to try a new and much more flexible paper submission policy.

    -- Authors may submit papers and tutorial / workshop proposals ANYTIME from now (20 September) to the final date (1-March-2008). If because of the weekend you cannot submit the full paper / workshop / panel on March 1, please send an abstract only to the Chair with the subject line of:


    and you may automatically recieve an extension to include Monday, March 3.

    -- the Paper Referees will either give a commitment of "Yes, your paper or tutorial proposal or workshop proposal is accepted" or "no, and the reason why is..." within two weeks of electronic submission. If necessary, we can supply both electronic and physical-paper approval documentation should it be necessary for getting a US entry visa approved.

    -- Any author may use this early submit / early decision method; the Referees may also provide additional feedback (both for accepted and declined papers). After all, the real goal is to have the best conference papers possible at the conference.

    In this way, we can provide a much longer time window for overseas authors to get their US entry visas; we apologize for any inconvenience this may have caused in the past.

    What to submit for papers, tutorials, and workshops

    -- Submissions for standard papers should be the paper plus any optional supporting materials.

    -- Submissions for tutorials (basically one to two hour short courses) should be the proposed syllabus, a one-paragraph bio of the volunteering lecturer, and supporting materials. If a topic is long, split it into two parts (i.e. Frobbing I and II).

    -- Submissions for workshop topics (hour-or-longer discussion groups) should be a one-paragraph explanation of the topic, plus a one-paragraph bio of the volunteering workshop moderator.

    As before, please send conference submissions (and any other questions) to the conference chair, You should expect a confirmation within 1 business day, and referee's results within two weeks.

    Topics for Papers, Workshops, and Tutorials

    Valid topics for 2008 include not just plain spam, but "other cybercrimes" such as phishing, IM spam, SMS spam, MMORPG spam, blog spam, trackback spam, photo spam, stock pump-and-dumps, email con games, exploit marketing, zombie bots and bot armies, setting up antispam systems, and antispam countermeasures including hardware, software, wetware, and blue-ware (i.e. employing the police).

    The common thread remains the same - dealing with undesired and unsolicited electronic communications; that's the center point of this conference and proposals should relate to that.

    Paper awards

    We will again have an audience vote for the best paper award, and if at least two student-authored papers are presented, a best student paper award.

    Paper Submission

    Anyone doing work in the spam, anti-spam, or other related cybercrimes is welcome to submit their results; we don't discriminate between academic, corporate, or private researchers; everyone competes on an equal footing. Heck, if you're a spammer and you want to talk, that's OK; submit a proposal and we'll give it a fair look.

    Paper Judging Criteria

    Papers should be technical in nature, and should give enough information so that a reasonable coder could reproduce your work and see if it works for them as well (science requires a reproducible experiment; without reproducible results we are not doing science).

    Statistical analysis papers and papers that use unique facilities to gather that information are welcome as well; assume that the reasonable coder could have access to big-ISP facilities, computation, or bandwidth if your method requires it.

    Paper Anti-Criteria

    Papers should not be sales pitches or "high level overviews"; lack of enough technical information to reproduce your results will probably cause your paper to be rejected. The one big exception to this is that text corpora of private communications are not expected to be published.

    Paper Mechanical Criteria

    As our conference publication method is strictly electronic, we have no rigid limits on page count. Make your paper exactly as long as it needs to be, and not a word longer; if you bore the judges, that will influence their scoring. We can accept papers in any of these formats: PDF, Microsoft Word, OpenOffice, and plain ASCII text papers. Additionally, we can accept "enhanced" submissions, so include proposed slides (even drafts) if you feel it will help the judges in their decision. We don't require 'blinded' papers, nor perfect final drafts until the time of the conference.

    Please name your paper's filenames by the last name of the author who should be "first contact" on the paper, and by a central keyword; for example, if you were Max Brooks and writing about zombie armies, then:

    would be a good submission filename. Please do NOT name your paper something like "MIT_SC_2008.pdf" because everyone does that and we end up with dozens of papers all with the same name and that never works out well.

    Paper republication/replication and copyright

    By submitting a paper to the MIT Spam Conference 2008, you grant a universal right for your paper to be electronically reproduced by any interested party for the purposes of learning and research.


    Every paper will be read by at least three referees; referees may recuse themselves for any reason on a paper and do not need to explain their reasons for recusement. If time allows, referees may give improvement feedback to paper authors; we want the best conference papers possible and referee comments are valuable insights. Referee feedback is by default anonymous; however a referee may choose to reveal their identity with their feedback on improvement; like recusement, referees may exercise this option without explanation.

    Conference Sponsorships

    As the MIT Spam Conferences are "free for all interested parties to attend" events, we actively solicit sponsorships. Sponsorships are orthogonal to paper acceptance; sponsoring (or not) will not affect whether a paper is accepted (or not). What sponsorship *does* get an organization is casual, at-conference access to attendees to establish personal relationships (typically tables right next to the coffee and donuts); in the past this has led to both hirings and business partnerships. Sponsorships are based on a "not to exceed" pledging system, and sponsorship pledge sizes are scaled to the size of the organization.

    Sponsors may also hold off-campus events; these events are "associated with but not part of" the MIT Spam Conference. Please contact the chair if your organization would like to do something like this. This year, COMCAST will be hosting an off-site reception (within easy walking distance) at 6 PM.

    Sponsorships are used to pay for instructional media, refreshments, and keynote speakers; the conference staff is all volunteer.


    If you can help run the conference, please let us know! We can use just about everything from simple "I know where the good restarants in Cambridge are" to paper referees. We especially can use people who know how to run AV equipment and can get video with good audio up onto YouTube in near-real-time.

    The 2007 Conference

    Thank you all for a great conference, especially the paper referees (who shall remain, as always, secret) and the ConCom members Lenny Foner and Bryt Bradley.

    We would like to thank MIT-CSAIL and MIT itself for allowing us the use of MIT facilities to hold this conference, and Mitsubishi Electric Research and Reputation Technologies for their financial sponsorship. The "Best Paper" award (by audience vote) goes to Ken Simpson and Stas Bekman of MailChannels for their paper on paper on SMTP Multiplex routing. In a moment of sheer irony, the trophy (a suitably decorated can of Genuine Hormel Spam ) was awarded to Ken, who then revealed that he is a long-time vegetarian.

    Papers and Proceedings

    The MIT Spam Conference proceedings for 2006 and 2007 are freely downloadable .iso CDROM images, which can be burned directly to CDROM for archive purposes; there are also gzipped tar archives available, if you're looking to read on a screen instead of burning directly to disk.
    If you are burning CD's and you have a couple extra, please burn two or three copies- one for yourself, one to be added to your department library, and one to be added to your university or institute library.

    MIT Spam Conference 2007 Complete Papers and Proceedings (.iso CDROM image)

    MIT Spam Conference 2007 Complete Papers and Proceedings (gzipped tar archive)

    MIT Spam Conference 2006 Complete Papers and Proceedings (.iso CDROM image)

    MIT Spam Conference 2006 Complete Papers and Proceedings (gzipped tar archive)

    Conference Videos

    Thanks to Rob Targosz of McAfee (who works on the McAfee SiteAdvisor system we have near-realtime YouTube videos. To find the videos, click on the link or serch YouTube for "spamconference" and "spam conference". The short videos below are these videos.

    Invited Topics
    9:35 Rich Segal: The Realtime Spam Filter Challenge realtime video
    9:45 Jessica Baumgart: Blog Spam realtime video
    10:15 Amanda Watlington: Search Engine Spam realtime video_1 realtime video_2 realtime video_3
    10:45 Coffee and Donuts I
    Considering the Source
    11:00 Alberto Trevino: Relays and Header Analysis Revisited realtime video_1 realtime video_2
    11:20 Alberto Mujica: Reputation Management for Email realtime video_1 realtime video_2
    11:40 Joseph McIsaac: SPF and Symmetric DNS realtime video_1 realtime video_2
    12:00 Lunch (on your own)
    Working the Text
    13:30 Aaaron Emigh: Automatically Detecting Textual Blog Spam realtime video_1 realtime video_2
    13:50 Catalin Cosoi: Combining antispam filters realtime video_2 realtime video_2
    14:10 Manuel Martin-Merino / Angela Blanco: Ensembles of SVM filters realtime video_2 realtime video_2
    14:30 Coffee and Donuts II

    Thinking Outside the text box
    15:00 Tobias Eggendorfer: Tarpit simulation realtime video_2 realtime video_2
    15:20 Drugge/Beckman/Simpson: SMTP Multiplex throttling realtime video_2 realtime video_2
    15:40 Fumero/Biggio: Detecting image spam realtime video_1 video_
    16:00 Ken Dallmeyer: ELIZAing the Spammers realtime video_1 realtime video_2
    ~16:20-onward Roundtable Rants / Late Breaking (all interested)
    and adjourn to Informal Discussions- that is, self-organized dinners.

    Due to transportation difficulties, the following paper was accepted to MIT SC 2007, but was not able to be presented:
    Nouman Azam: Feature selection and Latent Semantic Indexing
    . This paper is in the Proceedings CDROM (above).

    Extra Information on Spam Conferences

    For other Spam-Conference-like events, go to . In particular, the Messaging Anti-Abuse Working Group can be found at

    MIT Spam Conference 2007 Sponsors

    Contacting Us

    Conference Chair:
    William Yerazunis
    Senior Research Scientist
    Mitsubishi Electric Research Laboratories
    Cambridge, MA 02139
    Bio page
    Britton Bradley
    Last modified: Mon Apr 13 09:23:31 EDT 2009